AI Generated / Assisted Music

I tried the duo thing again with another song and it did not work.

I copied the lyrics from the one that did work and it worked again.

I know its hit and miss but its odd that it would do the same song 4 times and it works but a new one does not.:unsure:
 
The AI will try to make the song sound consistent so in its mind using a different voice probably breaks the consistency.
I suppose you had the right elements for it to work in that one song, but AI is unpredictable. It's also for this reason why you can't easily blend certain genres, eg, you can't (easily) start with a mellow jazz song which turns into a fierce death metal song by the end.
There is no conclusive, sure-fire way to force it.
It's a bit easier to do with Udio because they give you more control but that sometimes makes the end result seem more disjointed, and you don't necessarily get what you want. This is also why traditional music production skills are still useful even in an AI music workflow.
 
Suno and Udio are constantly one-upping each other... it's good for us who use their services. Suno recently released the capability to split instrumentals and vocals... and now Udio releases a new model, stem splitting and a lot more. I can't wait to try it out.

www.udio.com

Introducing v1.5

Udio v1.5 is our most advanced music model yet, with a host of improvements over our v1 model, including: • Improved Audio Quality • Key control • Improved global language results
www.udio.com
 
OnlyOneKenobi said:
Suno and Udio are constantly one-upping each other... it's good for us who use their services. Suno recently released the capability to split instrumentals and vocals... and now Udio releases a new model, stem splitting and a lot more. I can't wait to try it out.

www.udio.com

Introducing v1.5

Udio v1.5 is our most advanced music model yet, with a host of improvements over our v1 model, including: • Improved Audio Quality • Key control • Improved global language results
www.udio.com
Click to expand...
Ok. Udio's "Remix upload" feature is freaking awesome - so I can now take a Suno track, even if it has sound quality issues, chop it up into parts in a DAW, remix it in Udio - which seems like it clean it up nicely, and then have a CD quality AI generated song - with the composition of Suno and the quality of Udio. Dang. It takes a bit of work, but wow.
 
TribbleZA said:
Click to expand...
As far as stem separation is concerned, both tools use the same methods as most other "AI stem separation" services offer. It's not really what users wanted (those who requested it, wanted individually generated tracks for each instrument and then combined into a final song, like one would do in a studio) - but I understand why Suno and Udio don't or can't offer this at this point. I have my own stem separation workflow which is the same / better as Suno and Udio's methods.
 
OnlyOneKenobi said:
As far as stem separation is concerned, both tools use the same methods as most other "AI stem separation" services offer. It's not really what users wanted (those who requested it, wanted individually generated tracks for each instrument and then combined into a final song, like one would do in a studio) - but I understand why Suno and Udio don't or can't offer this at this point. I have my own stem separation workflow which is the same / better as Suno and Udio's methods.
Click to expand...
That is really cool. I think these AIs are nice and fun but I still believe humans need to be involved. There is something special about how humans combine sounds.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X